He's saying the switch is "up" but the SFP's stop communicating until a
reboot.
I assumed fiber SFP's. Maybe in a brownout the SFP dies first because
it's farther from the power supply than other components in the switch.
------ Original Message ------
From: "Chuck McCown" <[email protected]>
To: [email protected]
Sent: 8/23/2017 2:03:52 PM
Subject: Re: [AFMUG] Brown-outs, Power fluctuations lock up Mikrotik CRS
units
Glitches could come in through the ethernet side too.
-----Original Message----- From: Sterling Jacobson
Sent: Wednesday, August 23, 2017 11:59 AM
To: [email protected]
Subject: Re: [AFMUG] Brown-outs,Power fluctuations lock up Mikrotik CRS
units
I had about 10 of these units suffer from the disconnect problem this
morning.
Some were powered solely from DC wall plug directly to round connector.
Some were powered solely from POE injector from included DC wall plug
power supply.
Some were actually powered by both direct DC round plug AND POE from
the Ethernet/Ether1 port.
So in essence I've ruled all of those combos out as individual
immunity/culprits.
Some of these were plugged in to small UPS units from APC or similar
brand.
Which I know don't do line conditioning, so that doesn't surprise me
there.
-----Original Message-----
From: Af [mailto:[email protected]] On Behalf Of Adam Moffett
Sent: Wednesday, August 23, 2017 11:56 AM
To: [email protected]
Subject: Re: [AFMUG] Brown-outs, Power fluctuations lock up Mikrotik
CRS units
We have at least 50 CRS and I've never seen that.
We aren't using the wall warts though. These are almost all on Traco
48v power supplies plus a 48 to 24 converter. So I have a second layer
of conversion between the device and the power. That's just an
observation and I don't know that it ought to make a difference.
------ Original Message ------
From: "Sterling Jacobson" <[email protected]>
To: "[email protected]" <[email protected]>
Sent: 8/23/2017 1:39:43 PM
Subject: [AFMUG] Brown-outs, Power fluctuations lock up Mikrotik CRS
units
I get power glitches from time to time in our city and it never fails
to cause one or more of our CRS units to glitch.
They appear powered and running, but SFP+ modules don't communicate
until a power cycle.
Anyone else see this?
And generally, I don't understand how variations in 110v AC get
through
the DC power wart and cause problems to the CRS.
I mean, shouldn't the conversion from AC to DC inherently clean up
some
of the crap of the AC power?
Or is this a function of the power dipping low enough for brief time,
that the DC also goes almost out then surges back?
The CRS handles a wide range of DC input voltage, so I thought it
would
be more immune to fluctuations of DC voltage too.
Is that what is causing problems with the CRS maybe?
